Analysis combining correlated glaucoma traits identifies five new risk loci for open-angle glaucoma The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citation Gharahkhani, P., K. P. Burdon, J. N. Cooke Bailey, A. W. Hewitt, M. H. Law, L. R. Pasquale, J. H. Kang, et al. 2018. “Analysis combining correlated glaucoma traits identifies five new risk loci for open- angle glaucoma.” Scientific Reports 8 (1): 3124. doi:10.1038/ s41598-018-20435-9. http://dx.doi.org/10.1038/s41598-018-20435-9. Published Version doi:10.1038/s41598-018-20435-9 Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:35014999 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA www.nature.com/scientificreports OPEN Analysis combining correlated glaucoma traits identifes fve new risk loci for open-angle glaucoma Received: 23 August 2017 Puya Gharahkhani 1, Kathryn P. Burdon 2, Jessica N. Cooke Bailey 3, Alex W. Hewitt 2, Accepted: 18 January 2018 Matthew H. Law 1, Louis R. Pasquale 4,5, Jae H. Kang 5, Jonathan L. Haines3, Published: xx xx xxxx Emmanuelle Souzeau 6, Tiger Zhou6, Owen M. Siggs 6, John Landers6, Mona Awadalla6, Shiwani Sharma6, Richard A. Mills6, Bronwyn Ridge6, David Lynn7, Robert Casson8, Stuart L. Graham9, Ivan Goldberg10, Andrew White10,11, Paul R. Healey10,11, John Grigg10, Mitchell Lawlor10, Paul Mitchell11, Jonathan Ruddle12, Michael Coote12, Mark Walland12, Stephen Best13, Andrea Vincent13, Jesse Gale14, Graham RadfordSmith1,15, David C. Whiteman1, Grant W. Montgomery1,16, Nicholas G. Martin1, David A Mackey2,17, Janey L. Wiggs4, Stuart MacGregor1, Jamie E. Craig6 & The NEIGHBORHOOD consortium* Open-angle glaucoma (OAG) is a major cause of blindness worldwide. To identify new risk loci for OAG, we performed a genome-wide association study in 3,071 OAG cases and 6,750 unscreened controls, and meta-analysed the results with GWAS data for intraocular pressure (IOP) and optic disc parameters (the overall meta-analysis sample size varying between 32,000 to 48,000 participants), which are glaucoma-related traits. We identifed and independently validated four novel genome-wide signifcant associations within or near MYOF and CYP26A1, LINC02052 and CRYGS, LMX1B, and LMO7 using single variant tests, one additional locus (C9) using gene-based tests, and two genetic pathways - “response to fuid shear stress” and “abnormal retina morphology” - in pathway-based tests. Interestingly, some of the new risk loci contribute to risk of other genetically-correlated eye diseases including myopia and age-related macular degeneration. To our knowledge, this study is the frst integrative study to combine genetic data from OAG and its correlated traits to identify new risk variants and genetic pathways, highlighting the future potential of combining genetic data from genetically-correlated eye traits for the purpose of gene discovery and mapping. OAG is characterized by optic nerve damage and progressive loss of peripheral vision, with many patients remain- ing undiagnosed until severe irreversible vision loss has occurred1,2. OAG has a signifcant genetic component 1QIMR Berghofer Medical Research Institute, Brisbane, Queensland, Australia. 2University of Tasmania, Hobart, Tasmania, Australia. 3Population and Quantitative Health Sciences, Institute for Computational Biology, Case Western Reserve University School of Medicine, Cleveland, OH, USA. 4Department of Ophthalmology, Harvard Medical School, Massachusetts Eye and Ear Infirmary, Boston, MA, USA. 5Channing Division of Network Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA. 6Department of Ophthalmology, Flinders University, Adelaide, South Australia, Australia. 7South Australian Health & Medical Research Institute, School of Medicine, Flinders University, Adelaide, South Australia, Australia. 8South Australian Institute of Ophthalmology, University of Adelaide, Adelaide, South Australia, Australia. 9Ophthalmology and Vision Science, Macquarie University, Sydney, New South Wales, Australia. 10Department of Ophthalmology, University of Sydney, Sydney, Australia. 11Centre for Vision Research, The Westmead Institute for Medical Research, University of Sydney, Westmead, NSW, Australia. 12Centre for Eye Research Australia (CERA), University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Victoria, Australia. 13Department of Ophthalmology, University of Auckland, Auckland, New Zealand. 14Department of Ophthalmology, University of Otago, Dunedin, Otago, New Zealand. 15School of Medicine, University of Queensland, Herston Campus, Brisbane, QLD, Australia. 16Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia. 17Centre for Ophthalmology and Visual Science, Lions Eye Institute, University of Western Australia, Perth, Australia. *A comprehensive list of consortium members appears at the end of the paper. Correspondence and requests for materials should be addressed to P.G. (email: [email protected]) or J.E.C. (email: [email protected]) SCIENTIFIC REPORTS | (2018) 8:3124 | DOI:10.1038/s41598-018-20435-9 1 www.nature.com/scientificreports/ Meta-analysis Chr SNP Risk allele P-value Analysis heterogeneity P Nearest Genes 10 rs72815193 G 6.10 × 10−10 OAG + VCDR 0.31 MYOF and XRCC6P1 3 rs56962872 G 2.81 × 10−8 OAG + VCDR 0.51 LINC02052 and CRYGS 9 rs6478746 G 4.54 × 10−8 OAG + CA 0.44 LOC105376277 and LMX1B 1 rs148639588 T 3.53 × 10−8 OAG + CA 0.71 COL11A1 Table 1. Association results for the best SNPs within the genome-wide signifcant regions in meta-analyses of ANZRAG OAG and the endophenotypes. Efect sizes of these SNPs on OAG are presented in Table 2. OAG, open-angle glaucoma; CA, cup area; VCDR, vertical cup to disk ratio. ANZRAG NEIGHBORHOOD combined Efect Other Chr SNP allele allele OR SE P-value OR SE P-value OR SE P-value 10 rs4918865^ C G 1.149 0.03 1.89 × 10−5 1.086 0.04 0.01958 1.119 0.02 2.31 × 10−6 3 rs56962872 A G 0.862 0.04 2.91 × 10−5 0.892 0.04 0.002324 0.876 0.03 3.03 × 10−7 9 rs6478746 A G 0.853 0.04 1.09 × 10−5 0.909 0.04 0.01231 0.879 0.03 9.10 × 10−7 13 rs9530458 T C 1.158 0.03 4.50 × 10−6 1.138 0.03 0.00018 1.148 0.02 3.45 × 10−9 Table 2. GWAS statistics for the new OAG loci in ANZRAG (the discovery OAG set), NEIGHBORHOOD (the replication OAG set), and combined (fxed-efect meta-analysis). ^rs4918865 in high LD with rs72815193, LD r2 = 0.93; OR, odds ratio; SE, standard error of regression coefcent. with a relative risk of over 9 in frst-degree relatives of afected individuals compared to relatives of unafected people3. Our previous genome-wide association studies (GWAS) have reproducibly identifed several risk loci for OAG including TMCO1, CDKN2B-AS1, SIX6, CAV1, CAV2, ABCA1, AFAP1, GMDS, ARHGEF12, TXNRD2, ATXN2, and FOXC14–9. However, the majority of the genetic variance contributing to OAG remains unexplained, emphasizing that further studies to identify additional risk loci for OAG are required in order to make genetic risk prediction more clinically useful. Optic disk parameters including cup area (CA; the central area), disc area (DA; the total area of optic disc including cup area and the surrounding area containing axons of the retinal ganglion cells), and vertical cup-disc ratio (VCDR; the ratio of the vertical diameter of cup area to the vertical diameter of the optic disc) are key measurements used to assess OAG diagnosis and progression10. Elevated intraocular pressure (IOP) is the major known risk factor for OAG2. We refer to CA, DA, VCDR, and IOP as OAG endophenotypes or quantitative traits. Tere are high genetic correlations between these quantitative traits and OAG, with several of the already known risk loci for these traits overlapping with each other and with OAG loci, demonstrating their utility as endo- phenotypes11–14. Tese fndings suggest that combining genetic data from OAG and its endophenotypes has the potential to increase the probability of identifying genetic variants that are common between traits, thus enabling the extraction of greater genetic power from valuable disease cohorts. In this study, we sought to identify additional risk loci contributing to OAG susceptibility by (1) increasing the sample size for OAG, (2) combining GWAS data from OAG and its endophenotypes in order to increase our statistical power to identify new risk loci for OAG, and (3) applying gene and pathway based approaches. Results In total, 3,071 OAG cases from the Australian & New Zealand Registry of Advanced Glaucoma (ANZRAG) obtained in three phases of data collection, and 6,750 unscreened controls of European descent were used as the GWAS discovery dataset in this study (Supplementary Table 1). Five loci were associated with OAG at genome-wide signifcance level in the meta-analysis of GWAS results between the three phases of ANZRAG data (P < 5 × 10−8), including regions near or within CDKN2B-AS1, ABCA1, C14orf39 and SIX6, TMCO1, and ARHGEF12, all of which are now well established risk loci for OAG5–7,9. Manhattan and Q-Q plots are shown in Supplementary Figure 1. Genomic infation factor lambda was 1.006 for this analysis. Next, to increase the power of this study to identify new risk loci for OAG, we performed GWAS meta-analyses of ANZRAG OAG and each of the endophenotypes (CA, DA, VCDR, and IOP) that we obtained from our previous study11 (Supplementary Table 1). Before performing the meta-analyses, we confrmed validity of the endophenotypes for OAG by showing that there were signifcant genetic correlations between OAG and the endo- phenotypes (ranging between 20% and 47%, Supplementary Table 2; p ≤ 0.018) using the cross-trait bivariate LD score regression approach15.